Data Quality in Primary Care Electronic Medical Records in Manitoba
Background: Evaluation of primary care EMR data quality is crucial since data must be of high quality in order to maximize patient care and use databases for secondary purposes including improved chronic disease management. Completeness evaluates data for gaps that may limit it's ability to represent what it should. This study aims to evaluate the baseline problem list completeness for Manitoba primary care EMRs. Methods: We conducted a retrospective analysis of the QHR Accuro® EMR database within 9 salaried Winnipeg Regional Health Authority (WRHA} and 3 fee for service primary care clinics in Manitoba. Queries were designed in the Accuro® EMR query builder. Aggregates were used to calculate sensitivity as a measure of completeness. The seven chronic diseases evaluated include, hypertension, diabetes, hypothyroidism, asthma. COPD. CHF, and CAD. Only searchable, structured data with ICD-9 coding was assessed. The 12 clinic types were divided into four categories; teaching, access centres, community, and fee for service and mean completeness was calculated for each. One way AN OVA and post hoc contrast analyses were conducted to identify differences between salaried and fee for service clinics. Results: Fee for service clinics exhibited significantly lower problem list completeness rates than salaried clinics for hypothyroidism, asthma, COPD, and CAD. Sensitivities calculated for each disease were significantly worse than those reported from previous UK research. Conclusion: This study demonstrates the need for better understanding of data quality in Canada and improvements so that primary care data can be reliably used for secondary purposes.
data quality, electronic medical records